It is still early. Today it runs alongside Hasura, and Hasura still owns metadata authoring. The part we are focusing on first is compatibility on the GraphQL request path: metadata -> schema per role -> `/v1/graphql`.
The most useful feedback for us would be from people with real Hasura projects: schema diffs, metadata that does not work, behavioral differences, or queries that produce different results.
Happy to answer technical questions about the implementation, compatibility layer, or why we chose Go.